I wrote this for several reasons. Firstly, someone may find it useful. Secondly, in the final 18 months of my writing I was working full time and these methods helped. Finally, and probably the biggest reason, was that when I was nearing the end of my second year I attended a writing course for PhD students. I was struggling to write and thought the course would help.
The course was attended by PhD students from all disciplines, which made much of it irrelevant to those working in the sciences. It focused on enjoying writing, getting the most out of it and conveying your message. While I appreciate the idea of how it was structured, I didn't want this, I wanted actual practical advice. This was an opinion shared by many from the sciences who attended. In the end I never really received much advice in this area.
So with that in mind, and while it is still fresh in my mind, I have put it here in case it helps you.
Okay, so I started writing the thesis quite early, certainly a lot earlier than many of my colleagues. I was around two years into my PhD when I first began writing. However what was important was that I didn't actually begin with the intention of writing my thesis at this time. All I did was to decide that I wanted to use LaTeX so I created a template and added some results from an experiment. As I continued to perform more experiments and gather results, the more I had written in my LaTeX template. It didn't matter that the order was wrong, the structure of thesis changed many times, copy and paste is much easier than writing.
A blank page is the worst thing to have to deal with when you first start out. The enormity of the situation when you first begin is enough to make you do something else. Believe me, you will find something else to do.
So here is how I made a really good start, this will probably only take you a morning. I found that by simply fleshing out your thesis with chapters made a massive difference. It is quite likely that you will have looked at a couple of thesis already and there is a style that your field of research generally uses. Even if you don't know what chapters you will have, you certainly will have an introduction, background, methods, results and conclusions chapters. It doesn't matter if you change the structure at a later date. I found that since LaTeX creates the contents page, bibliography, titles and extra blank pages automatically, I suddenly had a 20 page thesis!
This can then be repeated for each chapter by fleshing out the sub-sections. Then within sub-sections you can make bullet points about what you need to talk about. You can keep repeating this until all you need to do is write the sentence.
Making figures and pretty graphs can be a lot of fun, however at the last minute they can be a nightmare. When deadlines start getting closer, I found that it was much harder to write in the evenings after doing a whole day of writing in the office. I would set myself goals to achieve each day and wouldn't be able to do it because I couldn't write at the same rate throughout the day. This would demotivate me as I missed my target.
To combat this I set more realistic targets of writing that could be achieved during the day along with creating a specific figure in the evening. So, after a day of writing I could work on a figure in the evening whilst watching the TV. This became much more motivating as I would add several pages in a day as a figure takes up a lot of space, and it's easier to write about something you have on the page. Having a more relaxed evening that is still productive helped with making me productive the following day.
In my final year of the PhD, I was working as a notetaker during term time. This involved attending around 15hours of lectures per week which allowed me to earn enough money for when the students left and I could no longer earn any money. Since lectures were spread out over the week, I would often find myself in a situtation where I could not work uninterrupted for more than an hour. I found that by making todo lists with tasks of varying time scales helped make myself always productive.
An example of this type of todo for one week;
I would do this on a Sunday evening and add to it and cross things off during the week. The following Sunday, I would transfer things that weren't done from last weeks list to the new one and add new things.
Now I don't mean the holiday part literally. However I did find that it was much easier to do certain things in particular places. In the office, I would write code, be organised, develop new algorithms and explore ideas. At home however, I could only ever write about things I had done (e.g. explain an algorithm or present some results) but I couldn't write about other peoples research. I found that if I could go away somewhere different for a weekend with all the papers and books I needed to explore and write about the current state of research, I would be much more productive than if I was in my office or at home.
So occasionally I would spend a long weekend at a friends house (when they weren't there) or at my Dad's caravan. This might sound like a bad idea, and if you're easily distracted, then it might be, however I found that by getting away, alone, in a place where there was nothing to do but write, I would come back with a sizable chunk of the thesis written.